基于L-M算法改进BP网络的信息对抗能力评估

在未来密集和复杂的电磁环境中,快速、客观地评估敌我双方的信息对抗能力具有重要的意义。目前通常采用的人工打分方法具有一定的主观性,且周期较长,难于满足战场瞬息万变的需求。提出了一种基于L evenberg-M arquardt算法改进的BP神经网络信息对抗能力评估方法,以某组信息对抗数据为训练数据,对改进BP神经网络进行训练,并进行了验证性的仿真试验。仿真结果表明:改进BP神经网络能客观有效地评估信息对抗能力,较大程度地提高了神经网络的收敛速度、缩短了评估时间。

Evaluation of Information Operations Improved BP Networks Based on L-M Algorithm

It is of the utmost importance to evaluate information operations quickly and objectively,the manual scoring method is the most common methods used for assessing it,however the methods are not quite objective,and also difficult to meet the needs of the rapidly changing battlefield conditions.So BP neural network improved by Levenberg-Marquard algorithm is put forward to evaluate the capacity of information operation.After the improved BP network is trained by the data of information operation from certain countries,and the simulation experiments are made.The results show that the improved BP neural network can effectively assess the capacity of information operations and it also can greatly improve the convergence rate of neural networks and shorten the evaluation time.